canonical: https://jentic.com/apis/agentcard.sh/agentcard

# Agentcard Sh Agentcard API

The Agentcard API verifies a user's identity from your own backend and connects their account through a send-code, verify-code, and consent flow. It runs KYC checks by accepting the front and back of a government ID plus submitted applicant information, then reports the verification status. Beyond identity, it issues and closes virtual cards for a verified member and places tracked purchases on their behalf. Calls are authenticated with a platform access token minted from your client ID and client secret.

## For AI agents

Verify user identities, run KYC document checks, issue and close virtual cards, and place tracked purchases on behalf of connected members.

## Scope

Does not handle payment settlement, accounting, or fraud scoring. Use for user identity verification, account connection, and member card issuance only.

## Capabilities

- Verify a user's identity by sending a one-time code and confirming it against their connected account
- Run KYC checks by uploading the front and back of a government ID and submitting applicant information
- Read a member's verification status and flow status before taking an action
- Issue a virtual card for a verified member and close it when it is no longer needed
- Place a purchase on a member's behalf and follow the order through its conversation and status
- Record a member's consent and refresh a connection before it expires

## Use cases

### Agent-Driven Identity Verification

An AI agent connecting to the Agentcard API through Jentic can verify a user without a human stepping in: it sends a one-time code, confirms the code the user returns, and reads back the verification status. Jentic injects the platform token at call time, so the agent follows the connect-and-verify flow without ever holding the credential in its prompt.

Example prompt: Call connect/start to send a verification code to a user, then call connect/verify with the code the user provides and confirm the connection succeeded

### KYC Document Onboarding

Collect a government ID and applicant details to meet know-your-customer requirements during onboarding. The Agentcard API accepts the front and back of the ID and the submitted information, runs the check, and exposes the verification status so your backend can gate access on a passing result. A test-mode simulate endpoint lets you rehearse each outcome before going live.

Example prompt: Upload the front and back of a user's ID, submit their KYC information, then poll the verification status until it resolves to a passing or failing outcome

### Virtual Card Issuance for Verified Members

Once a member is verified, issue them a virtual card and manage its lifecycle from the same backend. The Agentcard API lists a member's cards, creates a new card, reads a single card by its id, and closes a card when it is no longer needed, so your application controls spending instruments without a separate card platform.

Example prompt: Create a card for a verified member, read it back by its card_id to confirm it is active, and close it once the member no longer needs it

### Agentic Purchasing on a Member's Behalf

Let an agent place a purchase for a connected member and follow it to completion. The Agentcard API accepts a buy request, lists the merchants a purchase can be placed at, and exposes the purchase conversation and order status so the agent can report progress back to the member in real terms rather than guessing.

Example prompt: Submit a buy request for a member, then read the order by its order_id to report where the purchase currently stands

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| POST | `/api/v2/connect/start` | Send a verification code to a user |
| POST | `/api/v2/connect/verify` | Verify the code the user returns |
| POST | `/api/v2/kyc/information` | Submit applicant KYC information |
| GET | `/api/v2/kyc` | Read verification status |
| POST | `/api/v2/cards` | Issue a virtual card |
| GET | `/api/v2/cards/{card_id}` | Read a single card |
| POST | `/api/v2/cards/{card_id}/close` | Close a card |

## Key resources

- **Connect** — Send and verify one-time codes, record consent, and refresh a user's account connection
- **KYC** — Upload ID documents, submit applicant information, read verification status, and simulate outcomes in test mode
- **Cards** — List, create, read, and close virtual cards for a verified member
- **Buy** — Place purchases, list eligible merchants, and read purchase conversations and order status

## AI readiness

This API is usable in Jentic One now. Its AI-readiness score against Jentic's framework shows where it stands today and where improvements would make it even easier for agents to use.

- **Score:** 63 / 100
- **Maturity:** AI-Aware
- **Dimensions:**
  - Foundational Compliance: 58 / 100
  - Developer Experience & Jentic Compatibility: 66 / 100
  - AI-Readiness & Agent Experience: 52 / 100
  - Agent Usability: 94 / 100
  - Security: 60 / 100
  - AI Discoverability: 61 / 100
- **View full report:** https://jentic.com/apis/agentcard.sh/agentcard/scorecard
- **How the score is calculated:** https://docs.jentic.com/reference/api-readiness-framework/overview/
- **More about the dimensions:** https://docs.jentic.com/reference/api-readiness-framework/specification/#dimensional-model-overview

### Score it yourself

Every API in the directory is allowlisted, so you can re-score it with no key required.

- **Score your own API:** https://jentic.com/scorecard.md
- **Scoring CLI agent skill:** https://github.com/jentic/jentic-api-scorecard/blob/main/skills/jentic-api-scorecard/SKILL.md

```sh
npx @jentic/api-scorecard-cli score <openapi-url>
```

## Why Jentic

- **Setup:** Wiring the Agentcard API by hand means handling its bearer platform token minted from your client ID and secret, a separate user access token for member calls, and the connect-then-verify flow yourself. Through Jentic you install once, import Agentcard from the API Directory, store the token once, and your agent calls it.
- **Permission scoping:** You choose which Agentcard operations your agent may call, so identity checks like sending and verifying a code can be allowed while card issuance and closing stay out unless you add them. Card ids travel in the URL path (`/api/v2/cards/{card_id}`), so a rule can also pin card operations to a single card.
- **Credential handling:** Your Agentcard tokens are stored once, encrypted, by your own Jentic One instance and injected at execution time. They never enter the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'verify a user's identity' or 'issue a virtual card', and Jentic returns the matching Agentcard operation with its input schema so the agent calls the right endpoint without reading the reference docs.

## Related APIs

- **Onfido** — Onfido runs document and biometric identity verification as a dedicated KYC provider.
- **Veriff** — Veriff verifies government IDs and selfies for identity assurance.
- **Plaid** — Plaid connects a user's bank accounts and returns account and identity data.

## FAQ

### What authentication does the Agentcard API use?

The Agentcard API authenticates with bearer tokens per its OpenAPI spec. Your backend mints a platform access token from your client ID and client secret and sends it as a bearer token, and member-scoped calls use a separate user access token. Through Jentic the token is injected at call time, so it never sits in the agent's prompt or logs.

### Can I issue virtual cards for a member with the Agentcard API?

Yes. Once a member is verified you can list their cards, create a new card, read a single card by its card_id, and close a card when it is no longer needed. Card issuance is a distinct set of operations from the identity and connect flow, so you can grant an agent one without the other.

### What are the rate limits for the Agentcard API?

The OpenAPI spec does not specify rate limits. Check the Agentcard documentation at https://docs.agentcard.sh for current limits before running a high-volume verification or card workload.

### How do I verify a user's identity with the Agentcard API through Jentic?

Search Jentic for 'verify a user's identity' and it returns the connect operations with their input schema. Your agent calls connect/start to send a one-time code, connect/verify to confirm it, and reads the result from the verification status endpoint. To run it on your own infrastructure, install Jentic One from its GitHub repo.

### Can I limit what my agent is allowed to do with the Agentcard API?

Yes. Write a rule that allows only the connect and verify operations, so the agent can confirm a user's identity but cannot issue a card or place a purchase, and every action it takes is logged. Because card ids appear in the URL path, a rule can also pin card operations to a single card_id.

### Is there an Agentcard MCP server?

You don't need an MCP server to give your agent the Agentcard API. Jentic connects it directly from the API Directory: import it, store your credential once, and your agent calls the connect, KYC, card, and buy operations on demand without loading another server's tool definitions into its context.